ianmcook commented on a change in pull request #9999:
URL: https://github.com/apache/arrow/pull/9999#discussion_r624226610
##########
File path: r/tests/testthat/test-dplyr-mutate.R
##########
@@ -185,6 +164,238 @@ test_that("mutate with single value for recycling", {
)
})
+# similar to
https://github.com/tidyverse/dplyr/blob/master/tests/testthat/test-mutate.r#L12-L6
+test_that("rownames are preserved", {
+ skip("Row names are not preserved")
+ df <- data.frame(x = c(1, 2), row.names = c("a", "b"))
+ expect_dplyr_equal(input %>% mutate(y = c(3, 4)) %>% collect() %>%
rownames(), df)
+})
+
+# similar to
https://github.com/tidyverse/dplyr/blob/master/tests/testthat/test-mutate.r#L18-L29
+test_that("mutations are applied progressively", {
+ df <- tibble(x = 1)
+
+ expect_dplyr_equal(
+ input %>% mutate(y = x + 1, z = y + 1) %>% collect(),
+ df
+ )
+ expect_dplyr_equal(
+ input %>% mutate(x = x + 1, x = x + 1) %>% collect(),
+ df
+ )
+ expect_dplyr_equal(
+ input %>% mutate(y = x + 1, z = y + 1) %>% collect(),
+ df
+ )
+
+ df <- data.frame(x = 1, y = 2)
+ expect_equal(
+ df %>% Table$create() %>% mutate(x2 = x, x3 = x2 + 1) %>% collect(),
+ df %>% Table$create() %>% mutate(x2 = x + 0, x3 = x2 + 1) %>% collect()
+ )
+})
+
+# similar to
https://github.com/tidyverse/dplyr/blob/master/tests/testthat/test-mutate.r#L37-L54
+test_that("can remove variables with NULL (dplyr #462)", {
+ df <- tibble(x = 1:3, y = 1:3)
+ gf <- group_by(df, x)
+
+ expect_dplyr_equal(input %>% mutate(y = NULL) %>% collect(), df)
+ expect_dplyr_equal(input %>% mutate(y = NULL) %>% collect(), gf)
+
+ # even if it doesn't exist
+ expect_dplyr_equal(input %>% mutate(z = NULL) %>% collect(), df)
+ # or was just created
+ expect_dplyr_equal(input %>% mutate(z = rep(1, nrow(input)), z = NULL) %>%
collect(), df)
+
+ # regression test for https://github.com/tidyverse/dplyr/issues/4974
+ expect_dplyr_equal(
+ input %>% mutate(z = 1, x = NULL, y = NULL) %>% collect(),
+ data.frame(x = 1, y = 1)
+ )
+})
+
+# similar to
https://github.com/tidyverse/dplyr/blob/master/tests/testthat/test-mutate.r#L71-L75
+test_that("assignments don't overwrite variables (dplyr #315)", {
+ expect_dplyr_equal(
+ input %>% mutate(z = {x <- 10; x}) %>% collect(),
+ tibble(x = 1, y = 2)
+ )
+})
+
+# similar to
https://github.com/tidyverse/dplyr/blob/master/tests/testthat/test-mutate.r#L77-L81
+test_that("can mutate a data frame with zero columns and `NULL` column names",
{
+ df <- vctrs::new_data_frame(n = 2L)
+ colnames(df) <- NULL
+ expect_dplyr_equal(
+ input %>% mutate(x = c(1,2)) %>% collect(),
+ df
+ )
+})
+
+# similar to
https://github.com/tidyverse/dplyr/blob/master/tests/testthat/test-mutate.r#L102-L106
+test_that("mutate disambiguates NA and NaN (#1448)", {
+ expect_dplyr_equal(
+ input %>% mutate(y = x * 1) %>% select(y) %>% collect(),
+ tibble(x = c(1, NA, NaN))
+ )
+})
+
+# similar to
https://github.com/tidyverse/dplyr/blob/master/tests/testthat/test-mutate.r#L117-L127
+test_that("mutate handles data frame columns", {
+ expect_dplyr_equal(
+ input %>% mutate(new_col = data.frame(x = 1:3)) %>% select(new_col) %>%
collect(),
+ data.frame(x = 1:3)
+ )
+
+ # mutate() on grouped data not supported in Arrow; this will be pulling data
back into R
+ expect_warning(expect_dplyr_equal(
+ input %>%
+ group_by(x) %>%
+ mutate(new_col = x) %>%
+ ungroup() %>%
+ select(new_col) %>%
+ collect(),
+ data.frame(x = 1:3)
+ ), "not supported in Arrow")
+
+ skip("rowwise() is not (yet) implemented in Arrow")
+ expect_dplyr_equal(
+ input %>%
+ rowwise(x) %>%
+ mutate(new_col = x) %>%
+ ungroup() %>%
+ select(new_col) %>%
+ collect(),
+ data.frame(x = 1:3)
+ )
+})
+
+# similar to
https://github.com/tidyverse/dplyr/blob/master/tests/testthat/test-mutate.r#L129-L142
+test_that("unnamed data frames are automatically unspliced", {
+ # this works in arrow
+ expect_dplyr_equal(
+ input %>% mutate(tibble(b = 2)) %>% collect(),
+ tibble(a = 1)
+ )
+
+ # not this
+ expect_error(expect_dplyr_equal(
+ input %>% mutate(tibble(b = 2), tibble(b = 3)) %>% collect(),
+ tibble(a = 1)
+ ))
+
+ # nor this
+ expect_error(expect_dplyr_equal(
+ input %>% mutate(tibble(b = 2), c = b) %>% collect(),
+ tibble(a = 1)
+ ))
+})
+
+# similar to
https://github.com/tidyverse/dplyr/blob/master/tests/testthat/test-mutate.r#L144-L148
+test_that("named data frames are packed", {
Review comment:
I think this test ("named data frames are packed") should be removed;
it's not testing anything about arrow besides that the data is pulled into an R
data frame and the dplyr operation is performed on that
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
